C A N A L D E E X P E R I E N C I A S H I D R O D I N Á M I C A S, E L P A R D O

Publicación núm. 217

DUCTED PROPELLERS. A SOLUTION FOR BETTER

PROPULSION OF SHIPS. CALCULATIONS AND
PRACTICE
POR

H. HAIMOV
M. J. BOBO
J. VICARIO
J. DEL CORRAL

Ministerio
de Defensa

MADRID
MAYO 2010
DUCTED PROPELLERS. A SOLUTION FOR BETTER
PROPULSION OF SHIPS. CALCULATIONS AND
PRACTICE
POR

H. HAIMOV
M. J. BOBO
J. VICARIO
J. DEL CORRAL

Trabajo presentado en el 1st Symposium on Fishing Vessel Energy Efficiency. E-fishing. Vigo (Spain)

Mayo 2010.
 First International Symposium on Fishing Vessel Energy Efficiency
Ducted Propellers. A Solution for Better Propulsion of Ships.
Calculations and Practice.
H. Haimov1, Mª. J. Bobo2, J. Vicario3, J. del Corral3
1
Propulsors department, CEHIPAR, Madrid, Spain.
2
Projects department, CEHIPAR, Madrid, Spain.
3
CFD department, CEHIPAR, Madrid, Spain.
Abstract— After enumerating the advantages of applying ducted
propeller units on ships a short review of the hydrodynamic
calculations methods is presented. A short description of the
prediction calculation methods available in CEHIPAR is given
together with a validation exercise for the RANSE code, based on
comparison with model series and test for open water behaviour
of the propulsor. A real practical case explained has shown that
an integral approach combining calculation methods and model
tests is able to solve successfully the hydrodynamic problem of
the propulsion of a fishing vessel based on ducted propellers. The
ship fulfilled the requirements for low vibration levels and
showed better propulsive efficiency. The conclusions refer to the
particular role of the different numerical approaches in the
engineering practice.
Keywords – cavitation, CFD, ducted propeller, Euler solver,
model tests, propulsion, RANSE code, validation.
I. INTRODUCTION
In various occasions, our institution, “Canal de
Experiencias Hidrodinámicas de El Pardo” (CEHIPAR), has
been asked to solve the propulsion of a ship that would suffer
severe cavitation, vibrations and noise if the ship hull and the
propeller unit are not properly selected and designed.
Various solutions are possible and applied depending on the
peculiarities of the case, among them a carefully designed
wake adapted conventional propeller (see for example Haimov,
2008), or end-plate propellers (Pérez Gómez et al., 2006). Here
we will focus on another option – the well known two-
component propulsor called ducted propeller (DP), consisting
of propeller located inside a nozzle.
Since Kort nozzles (annular foils) have been introduced in
the 30ties of the past century many ships have been fitted with
ducted propellers, especially tugboats and trawlers
experiencing high loads on the propulsor – condition for
obtaining gain of propulsive efficiency.
The advantages of the ducted propellers can be summarized as:
This work has been partially sponsored by the European Commission, Grant
Agreement 234182 of the Seventh Framework Programme.
E-Fishing, Vigo, Spain, May 2010
 Increased thrust
 Protection of the propeller blades
 Greater hydrodynamic efficiency
 Reduction of cavitation, vibrations and noise
 Better manoeuvrability using azimuthing thrusters
 Safety in ice,
among others. Applied with success for relatively slow ships,
their principal shortcoming is that they are more easily fouled
than the open propellers in the presence of seaweed and debris.
The first advantage is characteristic for the accelerating
type DP and although better regulation of the pressure can be
obtained by using decelerating flow type nozzles the attractive
gain in hydrodynamic efficiency explains their major use.
A significant jump in the practical use of DP has been
observed after the development of a series of propellers in
nozzles, especially those developed in NSMB, Wageningen,
between 1954-1959, and more when regression polynomials
permitted to search for optimum DP of the Ka series
(Oosterveld, 1970), and the modified Kc series. This later
resulted equivalent in performance to the former, as shown by
(Yossifov, Zlatev and Staneva, 1984).
New types and shapes of nozzles appeared later and
nowadays the interest in DP is still high, focusing, even more
than on the efficiency, on the possibility of reduction of
cavitation, noise and vibrations and to the improvement of the
steering of the ships.
II. CALCULATIONS
The simplification of the complex 3-D flow permitted the
development of the first theories for calculating ducted
propellers appearing in the period of 1955-1962 in the works
of Dickmann & Weissinger, and Morgan (1962). When the
power of the computers permitted it, more elaborated methods blades boundaries, a narrow band of finer mesh was added
based on the calculation of axisymmetric inviscid vortex flows near the leading and trailing edges and the tip. A similar
have been developed in the 70ties by Lewis and Ryan (1972), approach was followed with the nozzle borders. The grid used
Gibson and Lewis (1973), Varsamov and Haimov (1978, in the calculations contains close to 2 million cells over a
1979, 1985) etc. The propeller has been modeled as an sector. Figure 1 shows the computational grid on the propeller
actuator disc interacting with the nozzle. In the 80ties Falcao and the nozzle:
de Campos (1983) made a broad contribution in the field,
including also the sheer flow characteristics when calculating
the DP in non-uniform flow. Actually, Baltazar and Falcao de
Campos (2009) reported an advanced panel method for the
prediction of the DP performance. A panel method has been
also developed by Kerwin, Kinnas, Lee and Shih (1987) and
extended to the unsteady case of cavitation predictions by Lee
and Kinnas (2006).
Since 1999 various steady RANSE applications for DP have
been published: Abdel-Maksoud and Heinke (1999), Sánchez-
Caja, Rautaheimo and Siikonen (2001), to mention the
pioneers. A reduction of the considerable computer time
necessary has been reported by Hoekstra (2006) simplifying
the propeller contribution, modeling it as an actuator disk.
Figure 1. Gridding of the propeller and the nozzle using ICEM.
The scale effect, the boundary layer on the duct, the tip
leakage vortex flow in the gap between the blades and the In order to validate the parameters of the computations a
nozzle inner surface, the interaction with a rudder, etc. are case of open water corresponding to a scaled model of ducted
typical problems having considerable viscous character and propeller Ka4-60-19A, P/D=1.14 from the Wageningen series
several publications are trying to advance the knowledge on has been chosen, model tested in CEHIPAR (Bobo and de la
this matter (see for example Kim, Peterson and Stern (2004) Rosa, 2004). The computations have been carried out on 64
and Sánchez-Caja, Pylkkanen and Sipila (2008)). bits AMD “Opteron 250” computer of 16 GB main storage.
Computational time of about 5 hours for one regime (advance)
A. The RANSE solver and its validation for ducted propellers was necessary to obtain converging results of precision 10-4.
The CFD code used for our calculations is the commercial The results as open-water curves of thrust, torque and
RANSE solver Ansys CFX® (2009) with the meshing tool efficiency, without any empiric corrections, are presented in
ICEM®. The Reynolds Averaged differential Navier-Stokes the Figure 2 below. The nomenclature of the figure follows the
equations are numerically solved by a finite volume technique, recommendations of the ITTC, 2008.
discretizing the computational domains.
From the available turbulence models, following the
positive experience of other authors (Abdel-Maksoud and Ka-19A Serie RANSE code M odel Test
Heinke, 2002), for example, we used the shear stress transport
equations (SST model) as in previous experience for open
propellers (Haimov, Terceño and Trejo, 2007). 0.58
Eta0
The definition of the computational domains is done taking
0.48
into account the presence of a stationary and rotational bodies. 10K Q
The nozzle is placed into an external stationary cylindrical
Kt,10Kq,Eta0

0.38 K TT
domain, and the propeller is inside the internal rotational
domain. Continuity is required at the border of the domains.
0.28
K TP
The calculations have been carried out in uniform inflow
specified as the boundary condition at the inlet, being the 0.18
K TD
outlet defined by a constant pressure. The axial symmetry of
the flow has been used by imposing symmetry conditions on a 0.08
sector accounting for the number of blades, thus reducing the
size of the numerical problem. On the surfaces of the rigid -0.02
0.20 0.30 0.40 0.50 0.60 0.70 0.80
bodies the non-slip boundary condition is imposed. Advan ce coefficient
The quality of the mesh is critical in CFD calculation using
RANSE methods. In this paper, a non-structured mesh of
Figure 2. Open water curves: calculations, series regressions and model test.
tetrahedrons was used, including prism layers in the boundary
layer zones. In order to obtain a better mesh definition in the
 Except in the zone of little interest and small forces of IV. PRACTICAL CASE
advances beyond the maximum efficiency, the agreement An example of the successful application of ducted
between the RANSE calculations, the model tests and series propeller on a ship is the propulsion solution for the fishing
data is very good. In that region the prediction of the nozzle research vessel FRV_EB. The main particulars of the ship and
thrust is excellent. For the propeller, the deviations of thrust its propulsion characteristics are shown in Table 1.
and torque do not exceed 7%. The efficiency is predicted even
better (within 4%). The computation of one case having TABLE 1: MAIN PARTICULARS OF THE SHIP
drastic general increase of the grid resolution resulted in Length overall 29.0 m
insignificant improvement of the results. Breadth, moulded 7.5 m
Displacement 249 T
B. The Euler Solver combined with Liftinig surface method
Block coefficient 0.4875
The axisymmetric Euler equations solver combined with Draught, moulded 2.6 m
body force presentation of the propeller calculated by vortex Type of propulsion engine Diesel
lattice method (Kinnas, Young, Lee, Gu and Natarajan, 2003) MCR Power 670 kW
has been used to compute the ducted propeller in the non- Nominal revolutions 1800 rpm
uniform wake inflow. This inviscid approach is applied here Reduction ratio 1:6
with the additional simplification for the influence of the Type of the definitive propulsor Ducted propeller
effective wake restricted only to its circumferentially mean Ship Speed 12 knots
axial component. An early version of the iterative solution
using programs GBFLOW-3X/MPUF-3A provided by a The model tests (Bobo, de la Rosa, Masip, Quereda,
Consortium on Cavitating Propulsors leaded by prof. S. Pangusión, 2005) revealed the appearance of developed sheet
Kinnas was also applied to evaluate the sheet cavitation on the cavitation on the blades of the conventional open propeller
blades, as done in multiple occasions before for open from the stock of CEHIPAR. Figure 4 shows the cavitation
propellers in non-uniform flow (Haimov, Valle, Baquero, patterns on the blades at different angular positions:
2002). The time necessary for the calculations of a typical case
is around 30 minutes on a HP Alpha work station. Cavitation Model Test No. 4661 OTI: 2233-CAV
Ref. tests: Self-propulsion No. 16594 Open water No. 16591
III. MODEL TESTS Hull No. XXXX Propulsor No. XXXX
Model hydrodynamic tests were carried out in the Draught, aft 2.6 m Type: FPP
Shaft immersion: 1.7 m Diameter: 1.8 m
CEHIPAR Towing tank and Cavitation Tunnel. The nozzle of Shaft inclination: 0.0 degr. P/D r=0.7 0.915 --
ducted propellers is treated as part of the propulsion unit. Wake model: Wire mesh
During open water and self-propulsion tests the thrust of the Full scale conditions: Model scale conditions:
nozzle is measured simultaneously with the thrust, torque and Revolutions: 314 rpm Scale: 6.75
Thrust coefficient: 0.222 Cavitation no. (σn) 2.60
rate of revolutions of the propeller. A set up of the propulsor Ship speed 12.2 kn Revolutions: 18 rps
behind the ship model is shown in Figure 3 below.

Figure 3. Afterbody of the ship model with the ducted propeller

The cavitation tests, modelling the pressure conditions,

follow the thrust identity method. The inflow to the DP is
usually simulated by a wire mesh reproducing the nominal
Figure 4. Cavitation diagram observations from model test of open propeller.
wake field (Figure 5) behind the ship model.
 The considerable cavitation on the blades is due essentially the model propeller blades inside the nozzle presented on
to the pronounced non-uniformity of the wake field at the Figure 10 showed clear reduction of the extent of cavitation –
propeller, as can be seen from the contour diagram shown in the objective of the application of the DP.
Figure 5 and obtained in the towing tank by wake survey test.

0.9
0.8

0.8

0.6
0.5

0.95
0.4

0.8
5
5

5
65

0. 5
0.

3
3
0.
55
0.7
0.2 0. 0.45

0.1
0.4 .5
0.25

5
0.05

0.05
0
0.7 0.7
25
0.1 0.
0.3
0.1

5
0.

0.1
05
5

0.4
5
0.0
0.05

0.05

0.0
5

0.3

0.2

0.05
05
0.
0.4

5
0.0
0.35

0.15
0.25

0.0
5
0.25
0.1

0.1
0.2

0.3
0.0

0.15
5

0.05

Figure 6. Gridding of the Euler solver domain

Figure 5. Contour lines of the nominal wake from wake survey test of open
propeller

The adoption of this open propeller would create

inacceptable, especially for this kind of ships, conditions for
vibrations and noise and there was no margin in schedule to
modify the stern ship forms – the main cause of this severe
non-uniformity of the wake field. Additionally, in trawling
regime, the loads on the propeller increase, reducing the
propulsive efficiency. All this suggested that possible
improvement could be obtained substituting the conventional
propeller with a ducted propeller. A decelerating nozzle would
permit the decrease of the cavitation, but the effect on the
propulsive efficiency could be negative. It was decided to try
an accelerating nozzle DP expecting a positive effect from the
homogenizing action of the nozzle and some gain in
propulsive efficiency.
Figure 7. Numerical prediction of sheet cavitation on open propeller blades
Due to the advanced stage of the project the design was
very limited in time and restricted just to the adapting and
fitting of a ducted propeller of the Wageningen Ka-19A series
in the given hull aperture.
To check the option of substituting the open propeller with
DP, MPUF-3A was run for the open propeller and in
combination with the Euler solver – for the ducted propeller.
The gridding used for the DP case is shown on Figure 6.
The results for the extension of cavitation are shown in
Figures 7 and 8, respectively. Note that the numerical
diagrams do not include tip vortex presentation. Clear
decrease of cavitation extension and volume for the propeller
in nozzle is observed.
Then the propeller and nozzle models have been
manufactured and tests carried out. Measurements of the
velocity field inside the duct without the propeller confirmed
the reduction of the non-uniformity of the nominal wake
inside the nozzle (Figure 9). Observations of the cavitation on Figure 8. Numerical prediction of sheet caviation on ducted propeller blades
 efficiency gain would be more pronounced if a possibility
existed to optimize the propulsor.

0.2

0.2
0.3
0.4
Finally, the full scale trials conducted by TSI to measure

5
0.4
the vibrations of the hull structure (Beltran, Galindo, Sánchez-

0.55

0.4
0.3
0.1.1

0.15
0.

0.1
0 5

5
0 0
Herrera and Pérez, 2006) obtained values inferior to 2 mm/s,

0.25

0.5 0.35
being significant the propulsor’s contribution to this very good

0.0
5
result of vibrations norms accomplishment.
V. CONCLUSIONS
0.1
0.1

0.0
0

0
5

An integral approach combining numerical predictions with

0.05

model tests permitted to solve satisfactorily the propulsion of a

ship fitted with ducted propeller.
For the time being model test measuring of the steady
forces is still required for precise predictions of the DP
Figure 9. Contour lines of the nominal wake from test inside the duct. performance, especially at the final design stage. Nevertheless,
the precision of the predictions based on RANSE code or
Cavitation Model Test No. 4655 OTI: 2233-CAV adjusted inviscid code is sufficient for the actual application
Ref. tests: Self-propulsion No. 16604 Open water No. 13255 and is a of valuable help during the design process.
Hull No. XXXX Propulsor No. XXXX The versatility of the inviscid tools, like lifting surface and
Draught, aft 2.6 m Type: Ducted Propeller panels, as well as Euler solvers, and their ability to estimate
Shaft immersion: 1.7 m Diameter: 1.7 m the unsteady effects and the cavitation in reasonable time
Shaft inclination: 0.0 degr. P/D r=0.7 1.140 --
Wake model: Wire mesh Nozzle: 19A, L/D=0.5
make them still very useful for engineering purposes. The
Full scale conditions: Model scale conditions: satisfactory precision of the predictions for the sheet cavitation
Revolutions: 308 rpm Scale: 6.75 on both, conventional and ducted propellers blades can be
Total thrust coefficient:0.246 Cavitation no (σn) 3.00 deduced comparing Figures 7 and 8 with Figures 4 and 10.
Ship speed 12.2 kn Revolutions: 18 rps
The RANSE solutions can predict with sufficient
engineering precision the steady forces on the duct and the
propeller. This is of interest for specific, non-series designs, but
require computational time not always compatible with the
time limitations of the practical projects. They are very good
tools to assess the problems related to details of the flow. The
analysis in non-uniform flow and the cavitation prediction are
still pending tasks for the RANSE methods of the near future.

ACKNOWLEDGMENT
The authors would like to thank the Technical Area of
CEHIPAR for carrying out the model tests and acknowledge
the management of CEHIPAR for the support of this work.
Thanks are also due to Dr. Amadeo García.

REFERENCES
[1] Abdel-Maksoud, M., Heinke, H.-J., “Investigation of viscous flow
around modern propulsion systems”, CFD’99, Ulsteinvik, 1999.
Figure 10. Cavitation diagram from model test of ducted propeller. [2] Abdel-Maksoud, M., Heinke, H.-J., “Scale effects on ducted propellers”,
24nd Symposium on Naval Hydrodynamics, Fukuoka, 2002.
The model tests also showed that the spots of cloud [3] ANSYS CFX v.12 Documentation. 2009.
cavitation detected in the tip region of the blades of the [4] Baltazar, J., Falcao de Campos, J.A.C., “On the modelling of the flow in
conventional propeller (Figure 4) do not appear in the ducted ducted propellers with a panel method.”, First Intern. Symposium on
propeller blades (Figure 10). Marine Propellers, Trondheim, 2009.
[5] Beltrán P., Galindo C., Sánchez-Herrera C., Pérez Fco.A., Silent fishing
An additional positive result obtained from the self- ships: Milestone of the small and medium size shipyards in Spain.
propulsion tests was the small gain in speed, due to a slight Lessons learned”. Ingeniería Naval, Febrero 2006, (In Spanish).
increase of the propulsive efficiency. This latter was due to the [6] Bobo Mª J., De La Rosa J-C, “ Selection of stock propeller and model
considerable increase of hull efficiency of DP compared with tests of a fishing research vessel for M.Cies Shipyards”, OTI 2233-1,
CEHIPAR, Nov. 2004. (In Spanish)
the open propeller. There is no doubt that the propulsive
[7] Bobo Mª J., De La Rosa J-C, Masip J., Quereda R., Pangusión L, “
Design of ducted propeller and model tests of a fishing research vessel
for M.Cies Shipyards”, OTI 2233-2, CEHIPAR, Jan-Apr. 2005. (In
Spanish).
[8] Falcao de Campos, J.A.C., “On the calculation of ducted propeller
performance in axisymmetric flows.”, PhD Thesis, Delft, 1983.
[9] Gibson, I., Lewis, R., “Ducted propeller analysis by surface vorticity and
acuator disk theory”, Symp. On Ducted Propellers, RINA, 1973.
[10] Haimov H., Valle J., Baquero A., “Implementation of CFD Tools for
Improvement of Ship Hydrodynamic Performance.” MARIND’2002, Varna
[11] Haimov, H., Terceño M., Trejo, I., “Use of commercial RANSE code for
open water propeller calculations.”, 10th NuTTS, Hamburg, 2007.
[12] Haimov, H. “Design of propellers of low vibroactivity and good
propulsive efficiency”, Ingeniería Naval, No.855, enero 2008, pp.24-25.
In Spanish
[13] Hoekstra, M., “A RANS-based analysis tool for ducted propeller
systems in open water condition.”, I.S.P., 53 (2006), 205-227.
[14] ITTC Symbols and terminology list, Version 2008.
[15] Kerwin, J., Kinnas, S., Lee, J., Shih, W., “A surface panel method for
the hydrodynamic analysis of ducted propellers“, Trans. SNAME,
vol.95., 1987.
[16] Kim, J., Peterson, E. and Stern, F., “RANS simulation of a tip-leakage
vortex flow on a ducted propeller marine propulsor.”, Journal of
Ship&Ocean Technology, March 2004.
[17] Kinnas S., Young Y., Lee H., Gu H., Natarajan S., “Prediction of
cavitating flow around single or two-component propulsors, ducted
propellers and rudders”, CFD 2003, London, 2003.
[18] Lee, H., Kinnas, S., “Prediction of cavitating performance of ducted
propellers”. Sixth Intrn.Symp. on Cavitation, Wageningen, 2006.
[19] Lewis R., Ryan P., “Surface vorticity theory for axisymmetric potential
flow past annular aerofoils and bodies of revolution with application to
ducted propellers and cowls”, Journal Mechanical Engineering Science,
vol.14, No.4, 1972.
[20] Morgan, W., “Theory of the annular airfoil and ducted propeller”, Forth
Symposium on Naval Hydrodynamics, 1962.
[21] Oosterveld, M.W.C., “Wake adapted ducted propellers”, Thesis of the
Technical University Delft, 1970. NSMB Publ. No.345.
[22] Pérez G., González-Adalid J., García A., Masip J., Quereda R.,
Pangusión L., Minguito E., Beltrán P., Galindo C., Pérez M., “ Full scale
comparison of a spuperferry performance fitted with both high-skewed
and CLT blades.”, WMTC, London, 2006.
[23] Sánchez-Caja, A., Rautaheimo, P., Siikonen, T., “Simulation of
incompressible viscous flow around a ducted propeller using a RANS
equations solver.”, 23rd Symposium on Naval Hydrodynamics, Val de
Reuil, 2001.
[24] Sánchez-Caja, A., Pylkkänen, J., Sipilä, P., “Simulation of the
incompressible viscous flow around a ducted propeller with rudders
using a RANSE solver”, 27rd Symposium on Naval Hydrodynamics,
Seoul, 2008.
[25] Varsamov K., Haimov A. , “Axisymmetric Potential Flow in Ducts” ,
Journal of Hydronautics , vol.12, No.2, 1978.
[26] Varsamov K., Haimov A. , “Axisymmetric Potential Flow about a
Multiply-Connected Body in Duct” , Proc. of 6th Conference on Fluid
Machinery, Vol.2, Budapest, 1979.
[27] Varsamov K., Haimov A., Minchev, D.“The use of varibale load
actuator disk in the design of ducted propellers.” , Proc. of
HYDROTURBO’85, Olomouc, 1985.
[28] Yosifov, K., Zlatev, Z., Staneva, A., “Optimum characteristics equations
for the K-J ducted propeller design charts”, 3rd Inetrn. Congress on
Marine Technology, Athens, 1984.